The Derwent World Patents Index (DWPI) database is well known for producing manually indexed records of specific and general chemical substances in patent documents. Processing more than a million new patent documents per year requires extraordinary efforts, and would be impossible without a robust information processing workflow. This paper will detail the role that text mining and natural language processing plays to assist with the Derwent editorial process, including the named entity recognition systems themselves, their place in the overall workflow, and the value that this indexing provides for searching patents by chemical and polymeric substances.

Augmenting manual curation of chemical patent information in the Derwent World Patents Index

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.